With the increasing penetration of renewable energy and the growing volatility of the power grid, the hybrid energy storage system has become an important means to enhance grid stability due to its multi-timescale response and energy complementarity characteristics. However, most traditional hybrid energy storage systems primarily rely on electrochemical and power-type storage, leaving a research gap for gravity energy storage as a novel physical storage form. This paper takes gravity energy storage as the core, combines it with battery energy storage devices, constructs a hybrid energy storage system optimization model considering the operational characteristics of gravity energy storage, and proposes a multi-objective collaborative control strategy aimed at improving system economy and response speed.
1. Introduction and Background
The significant increase in electricity demand driven by socioeconomic development and the proposal of the “dual carbon” goal have led to a yearly increase in the proportion of new energy generation. However, renewable energy sources represented by wind power and photovoltaics often exhibit characteristics such as intermittency, volatility, and anti-peaking behavior, which affect the power balance and frequency stability of the power system. Therefore, constructing large-capacity, long-duration energy storage systems can effectively address these issues. Among them, gravity energy storage technology, as an energy storage technology that transmits and converts gravitational potential energy, has advantages such as high safety, high efficiency, long lifespan, and short construction cycles, enabling its widespread application in multiple fields.
In traditional storage technologies, pumped hydro storage occupies a dominant position in the current market due to its mature technology and large-scale storage capacity, with global installed capacity exceeding 200 GW. However, its limitations are increasingly evident: first, it depends on special geographical conditions, and currently available development sites are gradually becoming saturated; second, its construction cycle is long and investment costs are high; third, it has certain impacts on the ecological environment, affecting river environments and ecosystems in some regions. Therefore, gravity energy storage, as an environmentally friendly and flexibly deployable new physical storage technology, has become a hot research direction internationally.
Under the background of accelerating global energy structure transformation toward low-carbonization, storage technology has become key supporting infrastructure for balancing supply and demand and enhancing grid resilience. Traditional single-type storage technologies (such as lithium-ion batteries, supercapacitors, flywheel storage, etc.) have their own advantages but also certain limitations. For example, gravity energy storage has high energy density but slower power response; supercapacitors have outstanding power density but limited storage capacity; and pumped hydro storage is constrained by geographical conditions. Therefore, hybrid storage technology has emerged, achieving precise matching of hybrid storage needs through synergistic complementarity of multiple types of storage media characteristics.
Currently, optimization of hybrid energy storage systems mostly focuses on well-developed storage technologies, and there are relatively few optimization strategies for hybrid storage systems incorporating gravity energy storage. However, based on the respective characteristics of gravity energy storage and batteries, such hybrid storage combinations have excellent development prospects. Nevertheless, the dynamic characteristic differences between gravity energy storage systems and batteries often make it difficult for such strategies to achieve global optimization.
By optimizing the cost of gravity energy storage and the capacity ratio between gravity energy storage system and battery storage, the full lifecycle cost of the hybrid system can be reduced. Utilizing the large-capacity and long-lifespan characteristics of the gravity energy storage system to undertake high-frequency, low-depth power fluctuation regulation during power generation can effectively reduce the number of battery cycles and significantly extend the service life of lithium batteries. When the gravity energy storage system stops due to failure or other reasons, the battery can provide emergency power support, thereby ensuring system power supply reliability.
According to estimates, if 10% of the national pumped hydro storage sites are replaced by gravity-battery hybrid energy storage systems, it would reduce ecological losses caused by reservoir inundation areas by approximately 150 square kilometers annually and lower carbon dioxide emissions by more than 8 million tons.
1.1 Current Status of Energy Storage Technologies
Currently, storage technologies are mainly showing a diversified and efficient development trend, with application scenarios expanding from the power system to industry, daily life, and many other fields. Among them, the battery industry, mainly lithium batteries, has a stable position, and emerging storage technologies such as sodium batteries, flow batteries, and hydrogen storage are accelerating toward commercialization. With national and local policy support and continuous technological updates in storage, storage technologies are moving toward scale and marketization, providing key support for the future global energy transition.
According to storage methods, commonly used storage methods can be roughly divided into electrochemical storage, thermal storage, electromagnetic storage, and physical storage. The different storage methods, characteristics, and applications are shown in the table below.
| Storage Type | Discharge Duration | Cycle Life (10,000 cycles) | Technical Features | Battery Type |
|---|---|---|---|---|
| Pumped Hydro Storage | 4–10 h | >5 | Mature technology, slow response, suitable for large scale | Energy-type |
| Flywheel Energy Storage | 15 s–15 min | 1–10 | High power density, high cost, fast response | Power-type |
| Compressed Air Storage | 1–2 h | >5 | Fast response, high cost, suitable for large scale | Hybrid energy/power type |
| Gravity Energy Storage | 2–10 h | >10 | Long lifespan, high efficiency | Hybrid energy/power type |
| Supercapacitor | 1–30 s | >5 | Fast response, high cost, long lifespan | Power-type |
| Superconducting Magnetic Storage | 2 s–5 min | >10 | Fast response, high cost, difficult maintenance | Power-type |
| Lead-acid Battery | 1 min–3 h | 0.2–0.4 | Low cost, mature technology, short lifespan | Hybrid energy/power type |
| Sodium-sulfur Battery | 1 min–5 h | <0.25 | High cost, high specific energy/power | Energy-type |
| Flow Battery | 1–10 h | 0.5–1 | High efficiency, environmentally friendly, deep discharge capable | Energy-type |
| Lithium Battery | 0.3–6 h | 0.2–1 | High energy density, high cost, pollution exists | Hybrid energy/power type |
1.2 Necessity of Hybrid Energy Storage System
To further develop storage technology and break through the technical performance boundary of single-type storage, hybrid energy storage systems that integrate the characteristics of multiple storage media have become another mainstream solution to enhance the comprehensive performance of storage systems. Typical hybrid modes mainly include:
(1) Power-type + energy-type storage combination. Power-type storage provides transient power support for the system, while energy-type storage undertakes medium-to-long-term energy dispatch. For the novel hybrid energy storage system based on modular multilevel converters, a hybrid storage synchronous control overall strategy has been proposed. In another study, batteries are used as auxiliary energy to store braking recovery energy embedded in supercapacitors with high power density, optimizing the problem that fuel cell output power is weak and cannot respond quickly to dynamic power loading. For the problem of low power density and slow dynamic response of battery-supercapacitor hybrid energy storage systems, a hybrid control strategy integrating robust variational physics-informed neural networks and snow ablation optimizer has been proposed, achieving multi-objective collaborative optimization of hydrogen consumption, load power management enhancement, power quality improvement, and dynamic voltage response improvement.
(2) Physical storage + electrochemical storage combination. This combination balances environmental adaptability while possessing excellent energy density advantages. The “flywheel + lithium battery” hybrid energy storage system operation control strategy improves the power supply capacity of the grid, enhances the frequency regulation performance of wind farms, and reduces the adverse impact of wind power fluctuations on the main frequency modulation of the grid. A hybrid storage smoothing-frequency regulation scheme based on flywheel and lithium battery has also been proposed, effectively improving the full lifecycle net income of the wind-storage combined system.
For the characteristics of gravity energy storage, its hybrid configuration with batteries has significant synergistic potential:
1) Time-scale complementarity: gravity energy storage is suitable for minute-to-hour level energy dispatch, such as peak shaving and valley filling; batteries can cover second-to-day level full-timescale regulation.
2) Economic optimization: the main cost of gravity energy storage is initial construction with low marginal cost, complementing the limited cycle life of batteries and reducing full lifecycle costs.
3) Improved system reliability: due to the heterogeneous characteristics of mechanical storage and electrochemical storage, the risk caused by common-mode faults can be effectively reduced, enhancing overall system robustness.
However, the engineering application of gravity-battery hybrid energy storage systems still faces many problems. For instance, the gravity energy storage system has inherent performance constraints; the mechanical structure design such as pulley group transmission ratio and mass distribution directly affects its dynamic response and efficiency, but existing studies using simplified static models lead to deviations between theoretical optimization results and actual operation. Moreover, the collaborative control of hybrid energy storage systems is also challenging; the energy allocation strategy needs to simultaneously adapt to the mechanical inertia delay of gravity storage and the electrochemical dynamic characteristics of batteries, making traditional rule-based control methods difficult to achieve multi-objective optimization.
2. Gravity Energy Storage Operational Characteristics and Structural Optimization
Gravity energy storage, as an emerging storage technology, has good development prospects. The gravity energy storage system mainly establishes a solid medium transport system to make it “flow” like water. The gravity storage system replaces water media in traditional pumped hydro storage with modular counterweight units, designing high-precision AGV transport vehicle groups, bridge crane matrices, multi-stage transmission chain mechanisms, and other intelligent unmanned operation systems. Taking slope-type gravity energy storage as an example, heavy blocks are transported between high and low positions on slope terrain. Through the synergistic action of the motor set and variable speed control system, heavy blocks can operate precisely within a specified speed range, simulating water flow characteristics.
The gravity energy storage system achieves energy management through conversion between potential energy and electrical energy. During the discharge process, heavy blocks at high positions slide down along the conveyor belt, driving the generator to complete the conversion from potential energy to mechanical energy and then to electrical energy. During the charging process, the motor drives the conveyor belt to transport heavy blocks from low positions to high positions, achieving energy circulation. Throughout the entire process, auxiliary systems such as laser radar positioning systems and IoT dispatch platforms are typically used to achieve closed-loop control of heavy block movement trajectories, ensuring temporal-spatial matching, reducing power tracking errors, and improving system stability.
2.1 Charging and Discharging Processes of Slope Gravity Energy Storage System
The charging process of the slope gravity energy storage system is a process of converting electrical energy into mechanical energy and finally into gravitational potential energy. During charging, the motor operates in motor mode, driving the transmission chain, thereby transporting heavy blocks from low positions to high positions for storage.
During the ascending process of heavy blocks, the transport speed of the transmission chain remains constant. The power consumed during the ascent of heavy blocks can be expressed as:
$$P_{L,in} = V_0 \times m g \sin(\theta) \quad (2-1)$$
where m is the mass of a single heavy block; g is the gravitational acceleration; and θ is the slope angle.
The power generated by the motor can be expressed as:
$$P_F = P_{L,in} + P_S = V_0 \times m g \sin(\theta) + P_S \quad (2-2)$$
where PS represents the additional losses in the gravity energy storage system, primarily consisting of mechanical losses caused by friction between components and electrical losses in the motor.
The charging efficiency of gravity storage is:
$$\eta_{g,in} = \frac{P_{L,in}}{P_{L,in} + P_S} = \frac{V_0 \times m g \sin(\theta)}{V_0 \times m g \sin(\theta) + P_S} = \frac{1}{1 + \frac{P_S}{V_0 \times m g \sin(\theta)}} \quad (2-3)$$
The discharge process of the gravity energy storage system involves placing heavy blocks at high positions on the conveyor chain and allowing them to slide down the slope under gravity, thereby driving the motor to generate electricity, realizing the conversion from gravitational potential energy to electrical energy. During the descent of heavy blocks, the motor brakes the conveyor chain while generating electricity to prevent the heavy blocks from falling too fast. The force analysis during the descent is:
$$f = F_n + F_f + F_s + F_z \quad (2-4)$$
$$m g = F_n \times \cos(\theta) + F_s + F_z + F_f \times \sin(\theta) \quad (2-5)$$
where Fn is the support force provided by the slope; Ff is the traction force provided by the motor; Fs is the force opposite to the direction of heavy block movement caused by mechanical losses and electrical losses; and Fz is the braking force of the motor. The discharge efficiency can thus be expressed as:
$$\eta_{g,out} = \frac{f \times \sin(\theta)}{…} \quad (2-6)$$
2.2 Multi-angle Segmented Control Method for Slope Gravity Energy Storage System Based on Auxiliary Storage Yard
For the gravity energy storage system, the energy release process exhibits discrete characteristics. Due to the limited descent speed of heavy blocks, the system response is slow and difficult to adapt to the rapid fluctuation demands of grid loads. Moreover, heavy blocks are large in size, and the terrain adaptability is insufficient. Slope-type gravity storage systems require continuous slope tracks to transport heavy blocks, imposing certain terrain conditions. They not only occupy large areas but also have strict slope requirements. To address these problems, this paper proposes a two-stage variable-angle slope structure for the slope of the gravity storage system and its corresponding control method. By combining slopes with different angles, it effectively solves the problems of poor flexibility, long reaction time, and difficulty in making timely adjustments when system loads change, while improving the terrain adaptability of the slope-type gravity storage system.
In addition to the proposed multi-angle slope segmentation control strategy, this paper also proposes a control strategy of adding auxiliary storage yards to assist the release of heavy blocks, effectively improving the output power accuracy of the slope-type gravity energy storage system, reducing discharge redundancy, and decreasing system response time.
The proposed dual-gradient slope coupling architecture with auxiliary storage yards is constructed as follows: the traditional single slope is decomposed into two adjustable slope segments, with a bidirectional buffer connection yard set between them. This yard serves both as the terminal storage compartment for heavy blocks transported from the top yard along the upper half slope and as the starting release compartment for heavy blocks transported along the lower half slope to the bottom yard. This structure not only improves energy transmission continuity but also enhances the adaptability range of the slope-type gravity storage system compared to single-slope structures.
Based on the instantaneous power of gravity, the instantaneous mechanical power generated by heavy blocks sliding along the upper and lower half slopes can be calculated as PF1 and PF2:
$$P_{F1} = m g V_0 \sin(\theta_1) \quad (2-7)$$
$$P_{F2} = m g V_0 \sin(\theta_2) \quad (2-8)$$
$$P_S = P_{F1} + P_{F2} \quad (2-9)$$
where V0 is the descent speed of heavy blocks along the slope; θ1 is the slope angle of the upper half slope; θ2 is the slope angle of the lower half slope; and PS is the sum of mechanical power generated when the same heavy blocks slide down each slope segment simultaneously.
To reduce energy redundancy in the gravity storage system, the instantaneous mechanical power generated when heavy blocks are released from the connection yard and the top yard should satisfy the load power demand Pt as much as possible without exceeding it:
$$X \times P_S < P_t < (X + 1) \times P_S \quad (2-10)$$
where X is the number of heavy blocks released from the top yard and connection yard to the upper and lower half slopes.
Since the descent process of released heavy blocks cannot be completed instantaneously, to cope with load fluctuations, the top yard and connection yard need to determine whether to release heavy blocks every Δt seconds. By comparing the energy generated by heavy blocks released from the top yard and connection yard within Δt seconds with the electrical energy consumed by the load, the system determines whether to continue releasing heavy blocks and the number to release:
$$W_t = \int_{t}^{t+\Delta t} P_t dt \quad (2-11)$$
$$W_{F1} = P_{F1} \times \Delta t \quad (2-12)$$
$$W_{F2} = P_{F2} \times \Delta t \quad (2-13)$$
where Wt is the electrical energy required by the grid within Δt seconds; WF1 is the electrical energy generated by one heavy block on the upper half slope within Δt; and WF2 is the electrical energy generated by one heavy block on the lower half slope within Δt.
Let the number of heavy blocks released from the top yard onto the slope be X1, and the number released from the connection yard be X2:
$$X_1 \times W_{F1} + X_2 \times W_{F2} < W_t < (X_1 + 1) \times W_{F1} + (X_2 + 1) \times W_{F2} \quad (2-14)$$
When there are already N1 and N2 heavy blocks on the upper and lower half slopes respectively released from the top yard and connection yard:
$$C_1 = X_1 – N_1 \quad (2-15)$$
$$C_2 = X_2 – N_2 \quad (2-16)$$
where C1 is the number of heavy blocks that need to be released from the top yard, and C2 is the number of heavy blocks that need to be released from the connection yard. When C1 ≤ 0, no heavy blocks are released from the top yard; when C1 > 0, C1 heavy blocks are released from the top yard. Similarly for the connection yard.
Based on the proposed gravity storage system structure, this paper adopts a standardized spatial topology architecture design. Modular auxiliary storage yard arrays are arranged uniformly on both sides of the dual-gradient slope, constructing a parametric three-dimensional coordinate system based on the slope gravity storage system. Heavy blocks are set as identical cubes, with the side length of the heavy block as the unit length of the coordinate system.
When positioned within the three-dimensional coordinate system, the gripper grabbing heavy blocks is perpendicular to the XOY plane and parallel to the ZOZ plane and YOZ plane. The initial position of the gripper faces the origin (0, 0, Z₀), and the gripper returns to the initial position after each grab. The horizontal movement speed of the gripper is V1, the vertical movement speed is V2. The horizontal distance for grabbing a heavy block at (X, Y, Z) is:
$$L_{xy} = 2 \times (x + y) \quad (2-17)$$
with constraints:
$$x \leq x_{max}, \quad y \leq y_{max} \quad (2-18)$$
where xmax is the maximum number of heavy blocks stacked along the X-axis, and ymax is the maximum number stacked along the Y-axis. The vertical distance is:
$$L_z = (Z_0 – z) \quad (2-19)$$
The time required for grabbing is given by:
$$t_{xyz} = \frac{L_{xy}}{V_1} + \frac{L_z}{V_2} \quad (2-20)$$
Taking the derivative of the power change curve P with respect to time gives the power change rate curve f. Segmenting the curve by unit time Δt, the average absolute value of the power change rate within unit time is:
$$D_t = \frac{1}{\Delta t} \int_{t_1}^{t_1 + \Delta t} f \, dt \quad (2-21)$$
$$t = t_{xyz,max} \quad (2-22)$$
Based on the time required to grab heavy blocks in the auxiliary storage yard and release them onto the slope, combined with the difference between the load power required by the system and the power generated by heavy blocks released from the top and connection yards, the number of auxiliary storage yard layers that need to release heavy blocks on the upper and lower half slopes can be calculated.
The time required for heavy blocks released from the auxiliary storage yard on the upper half slope to slide from the slope to the bottom is T1, and the average instantaneous power released during the process from grabbing to releasing to the connection yard is P1. For the lower half slope, the corresponding time is T2 and power is P2:
$$T_1 = \frac{n_1 \times \Delta h}{V \sin(\theta_1)} + t_1 \quad (2-23)$$
$$T_2 = \frac{n_2 \times \Delta h}{V \sin(\theta_2)} + t_2 \quad (2-24)$$
$$P_1 = \frac{W_g}{T_1 + \Delta t} = \frac{n_1 \times m g \times V \sin(\theta_1) \times \Delta h}{T_1 + \Delta t} \quad (2-25)$$
$$P_2 = \frac{W_g}{T_2 + \Delta t} = \frac{n_2 \times m g \times V \sin(\theta_2) \times \Delta h}{T_2 + \Delta t} \quad (2-26)$$
where n1 is the number of auxiliary storage yard layers on the upper half slope; n2 is the number of auxiliary storage yard layers on the lower half slope (both are integers greater than or equal to 0); Δh is the height interval between auxiliary storage yards; t1 is the grabbing time for heavy blocks in the upper half slope yard; t2 is the grabbing time for heavy blocks in the lower half slope yard; and V is the slope conveyor speed.
Since the heavy blocks released from the top yard and connection yard already satisfy most of the load demand, the auxiliary yards on the upper and lower half slopes only need to meet the remaining load demand, reducing energy loss and achieving precise power control. The output power of the auxiliary storage yard is:
$$P_y = P_t – N \times P_S \quad (2-27)$$
Since θ₁ is greater than θ₂, the function of the average instantaneous power generated by the upper and lower half slope yards is:
$$P_f(n) = \frac{n_1 \times m g \times V \sin(\theta_1) \times \Delta h}{T_1 + t_1} + \frac{n_2 \times m g \times V \sin(\theta_2) \times \Delta h}{T_2 + t_2} \quad (2-28)$$
The constraint is:
$$P_y < P_f(n) < P_{F1} + P_{F2} \quad (2-29)$$
To reduce the value of Tx and improve system response speed, it is necessary to keep n1 and n2 as close as possible. The expression for the number of auxiliary storage yard layers releasing heavy blocks is:
$$n_x = \frac{P_y \times t_1 \times V \sin(\theta_1)}{\Delta h \times (2 m g \times V \sin(\theta_1) – P_y)} \quad (2-30)$$
$$n_1 < n_x < n_2 < n_1 + 1 \quad (2-31)$$
2.3 Track Reversing Design for Slope Gravity Energy Storage System
Since the construction cost of auxiliary storage yards is relatively high, this paper designs a reversible slope track to replace the function of auxiliary storage yards. By controlling the direction of the track transporting heavy blocks on the slope, individual heavy blocks can be stored temporarily beside the slope track, avoiding waste caused by excessive heavy blocks on the track generating far more power than the load demand simultaneously. When needed, the heavy blocks can return to the track to continue the power generation process, thereby achieving more precise control of the output power of the slope gravity storage system.
First, a transport track for heavy blocks is established on the slope of the slope-type gravity energy storage system, and storage tracks for temporary storage of heavy blocks are built beside the transport track at intervals of Δl. The storage tracks are connected to the transport track through track switches, and the connection tracks are named connecting tracks. The storage track and transport track are connected by two connecting tracks in an “人” (inverted V) shape. When a heavy block trolley needs to travel to the storage track, the track switch on the transport track turns to the direction of the connecting track, allowing the heavy block trolley to travel along the upper connecting track to the storage track. When a heavy block trolley needs to return to the transport track, the track switch on the storage track turns to the connecting track, allowing the heavy block to travel along the lower connecting track to the transport track.
The instantaneous power generated when a single heavy block trolley slides down the track is:
$$P_F = m g V_0 \sin(\theta) \quad (2-32)$$
The top yard of the slope gravity storage system starts operation at time 0 and releases one heavy block trolley every Δt seconds. Starting from Δt/2, the track switch calculates the number of heavy blocks that should exist on the transport track every Δt seconds and switches the track switches to move heavy block trolleys from the transport track to the storage track or back. Each switch automatically returns to its original state after operation. The vertical descent speed of heavy block trolleys along the transport track equals that along the connecting track, and each storage track receives/releases only one heavy block trolley per determination. Therefore, the number of trolleys that should enter or leave the storage track every Δt seconds is:
$$n = X – N \quad (2-33)$$
where N is the number of heavy block trolleys on the transport track; when n is positive, n is the number of trolleys leaving the storage track to the transport track; when n is negative, the absolute value of n is the number of trolleys entering the storage track from the transport track.
Finally, the load power in the system is differentiated to obtain the load power change rate over time. The number of track switch layers that need to be activated is determined based on the speed of the load power change rate and the number of trolleys needing to change tracks. When the slope gravity storage system has three storage tracks, the bottom storage track is layer 1 and the top is layer J. Since the lower the storage track layer, the shorter the descent time of the released heavy block trolley on the slope, the faster the response speed of the gravity storage system. The rule for assigning heavy block trolleys to storage tracks at different layers is:
$$F = f_{max} – f_{min} \quad (2-34)$$
$$a = \begin{cases} 1, & f \in [f_{min}, f_{min} + F/3) \\ 2, & f \in [f_{min} + F/3, f_{min} + 2F/3) \\ 3, & f \in [f_{min} + 2F/3, f_{max}] \end{cases} \quad (2-35)$$
where a is the storage track layer; fmax is the maximum value of f; and fmin is the minimum value of f. When heavy block trolleys need to enter or leave the storage track, the track switches at the computed layer a and adjacent layers are selected. If the storage track at the selected layer is full/empty, adjacent storage tracks are selected instead.
2.4 Simulation Verification
Compared to traditional slope gravity storage systems, the segmented slope design increases construction cost but greatly improves the terrain adaptability of the gravity storage system. Since the segmentation reduces the single descent distance and descent time of heavy blocks, it effectively increases the release frequency of heavy blocks, thereby improving the control accuracy of input/output power during charging and discharging of the gravity energy storage system. The specific advantages and disadvantages are shown in the following table.
| Aspect | Segmented Slope Gravity Energy Storage Characteristics |
|---|---|
| Construction cost | Higher than traditional slope gravity storage; increases with the number of segments |
| Storage capacity | Slightly higher than traditional slope gravity storage due to connection yards that store additional blocks |
| Terrain adaptability | Significantly improves with increased slope segmentation |
| Cycle efficiency | Improved due to better power control accuracy after segmentation |
| Maintenance cost | Individual slope segment length is shortened; overall slope maintenance cost is lower, but connection yards increase overall cost |
| Response time | Faster response with more slope segments |
Considering that different slope lengths after segmentation have different performance improvement levels, this paper simulates multiple slope lengths and segmentation numbers, uniformly using a 30-degree slope angle with heavy blocks descending at 2 m/s. The simulation results for different slope lengths are shown in the table below.
| Slope Length | 10 m | 20 m | 30 m | 40 m |
|---|---|---|---|---|
| Deviation sample 1 | 0.1230 | 0.1646 | 0.1702 | 0.1973 |
| Deviation sample 2 | 0.1224 | 0.1648 | 0.1707 | 0.1949 |
| Deviation sample 3 | 0.1226 | 0.1650 | 0.1709 | 0.1951 |
| Deviation sample 4 | 0.1228 | 0.1643 | 0.1710 | 0.1952 |
| Deviation sample 5 | 0.1230 | 0.1635 | 0.1712 | 0.1954 |
| Deviation sample 6 | 0.1232 | 0.1637 | 0.1714 | 0.1956 |
| Deviation sample 7 | 0.1228 | 0.1639 | 0.1716 | 0.1958 |
| Deviation sample 8 | 0.1225 | 0.1641 | 0.1717 | 0.1960 |
| Deviation sample 9 | 0.1226 | 0.1642 | 0.1719 | 0.1961 |
| Deviation sample 10 | 0.1228 | 0.1644 | 0.1720 | 0.1963 |
| Deviation sample 11 | 0.1229 | 0.1645 | 0.1708 | 0.1964 |
| Average deviation | 0.1228 | 0.1643 | 0.1713 | 0.1958 |
From the table, it can be observed that the deviation between the output power of the slope gravity energy storage system and the load demand shows a negative correlation with slope length. When the slope length varies between 10 and 40 m, the average deviation rate increases from 12.2% to 19.6%. The standard deviations for slope lengths of 10 m, 20 m, 30 m, and 40 m are 0.08%, 0.15%, 0.18%, and 0.24% respectively, indicating that the fluctuation amplitude increases with slope length. In the gravity storage system, power deviation is mainly caused by load response delay and discrete output power, which proves that changing slope length can improve the response speed of the gravity storage system.
Based on the modular concept, the auxiliary storage yard technology provides a new upgrade path for the gravity storage system. After adding auxiliary storage yards on both sides of the slope, the system power regulation freedom increases. The separated structure ensures that the main slope parameters remain unchanged, reducing the cost of retrofitting existing gravity storage facilities. However, it should be noted that setting up auxiliary storage yards increases construction costs, and due to land area constraints, the capacity of auxiliary storage yards is much smaller than that of the main storage yard.
Simulation verification was conducted before and after adding auxiliary storage yards to the slope gravity storage system. The simulation parameters are shown in the table below.
| Parameter | Value |
|---|---|
| Height difference (m) | 20 |
| Slope angle (°) | 30 |
| Slope yard elevation per layer (m) | 2 |
| Slope length (m) | 40 |
| Block velocity on slope (m/s) | 5 |
| Block mass (t) | 0.1 |
| Slope yard layers | 9 |
| Equivalent vertical lift (m/s) | 2.5 |
| Gripper movement speed (m/s) | 0.5 |
| Total number of blocks | 8 |
After adding auxiliary storage yards, the new gravity storage device shows significant improvement in energy output accuracy compared to the traditional slope system. The power output curve of the gravity storage system with auxiliary storage yards improves the fitting degree with the target load curve by approximately 37%, and the precise power regulation mechanism effectively reduces energy redundancy losses, significantly improving the operational economy of the system.
In terms of structural optimization, the track reversing technology has the following benefits compared to auxiliary storage yards: first, the modular steering design can effectively reduce construction costs. Second, after adopting the reversing strategy, the impact force caused by heavy block inertia will be greatly reduced, improving equipment service life and operational safety. However, this technology requires a larger storage area and is therefore more suitable for medium-to-large gravity storage power stations.
3. Optimization Strategy for Gravity-Battery Hybrid Energy Storage System
In the current diversified hybrid storage architecture design, the battery-gravity storage combination has the following advantages over other types of hybrid storage:
(1) Time-frequency complementarity. Compared with electromagnetic storage such as flywheels and supercapacitors combined with gravity storage, the battery-gravity hybrid system is slightly weaker in millisecond response, but the energy duration expansion capability of the battery-gravity combination is outstanding: batteries can cover minute-to-hour short-period regulation, while gravity storage can support 4-12 hour long-period peak shaving, providing a relatively complete broadband regulation capability.
(2) Full lifecycle cost advantages. The levelized cost of storage of the battery-gravity system is lower than that of compressed air-gravity storage solutions. In addition, the long cycle life of lithium batteries and the long lifespan of gravity storage form a life-cycle synergy; gravity storage undertaking base load regulation reduces the cycle depth of lithium batteries, greatly extending battery service life.
(3) Strong terrain adaptability. Compared with compressed air-gravity storage systems that rely on underground caverns, or pure pumped hydro-gravity storage solutions requiring high head differences, the battery-gravity hybrid storage system is more flexible in deployment. Gravity storage modules can adapt to conventional slope terrain with small head differences, and battery packs can adopt containerized deployment to reduce land requirements.
(4) Environmental benefits. The full lifecycle carbon emissions of the battery-gravity hybrid storage system are lower than those of pumped hydro-gravity storage. The environmental friendliness of battery-gravity storage is also reflected in the fact that heavy blocks for gravity storage can use waste materials from construction or local raw materials. The coordinated scheduling of lithium batteries and gravity storage can greatly reduce the daily average cycle count of batteries, effectively mitigating electrolyte degradation. Moreover, this hybrid storage system does not require water resources, making it particularly suitable for arid regions.
The battery-gravity hybrid storage system can meet the stringent requirements of the new power system for flexible resources through a “second-level response + hour-level duration” operation mechanism, a “shallow charge/discharge + deep cycle” operation strategy, and a “solid-state + electrochemical” complementary design, becoming the preferred solution to support high-proportion renewable energy integration.
3.1 Charging and Discharging Efficiency Models
When the overall storage system is in the charging state, the new energy input power Pnew and load power Pl can be used to obtain the storage power Pstorage:
$$P_{storage} = P_{new} – P_l \quad (3-1)$$
Since the storage system is in the charging state, Pstorage > 0, and:
$$P_{new} > P_g + P_b > 0 \quad (3-2)$$
where Pg is the power of the gravity storage system and Pb is the battery power.
Let the charging efficiency of the gravity storage system be ηg,in, the actual power stored in the gravity storage system be Pg,in, and the discharging efficiency be ηg,out. The comprehensive efficiency of the gravity storage system is:
$$\eta_h = \eta_{g,in} \times \eta_{g,out} \quad (3-3)$$
When the gravity storage system is charging:
$$P_{g,in} = P_g \times \eta_{g,in} \quad (3-4)$$
When the gravity storage system is discharging:
$$P_{g,out} = \frac{P_g}{\eta_{g,out}} \quad (3-5)$$
For the battery, let the charging efficiency be ηb,in and the discharging efficiency be ηb,out:
$$\eta_b = \eta_{b,in} \times \eta_{b,out} \quad (3-6)$$
When the battery is charging (Pb > 0):
$$P_{b,in} = P_b \times \eta_{b,in} \quad (3-7)$$
$$P_{b,out} = 0 \quad (3-8)$$
When the battery is discharging (Pb < 0):
$$P_{b,out} = \frac{P_b}{\eta_{b,out}} \quad (3-9)$$
Since the energy released by gravity storage changes in a step-like manner, the power generated or stored when a single heavy block moves from bottom to top or from top to bottom is constant at Pg1:
$$P_{g,in} = n \times P_{g1} \quad (3-10)$$
where n is the number of heavy blocks (n > 0 and is an integer).
The charging efficiency function of the overall storage system is:
$$\eta = \frac{P_{storage,in}}{P_{storage}} \quad (3-11)$$
Since the overall storage system is charging, the function relating internal power changes to n and Pb is:
$$F(n, P_b) = \begin{cases} n \times P_{g1} \times \eta_{g,in} + P_b \times \eta_{b,in}, & P_b > 0 \\ n \times P_{g1} \times \eta_{g,in} + \frac{P_b}{\eta_{b,out}}, & P_b < 0 \end{cases} \quad (3-12)$$
Since the battery power and the storage power are related by:
$$P_b = P_{storage} – n \times P_{g1} \quad (3-13)$$
the binary function can be simplified to a unary function:
$$F(n) = \begin{cases} n \times P_{g1} \times \eta_{g,in} + (P_{storage} – n \times P_{g1}) \times \eta_{b,in}, & n < \frac{P_{storage}}{P_{g1}} \\ n \times P_{g1} \times \eta_{g,in} + \frac{P_{storage} – n \times P_{g1}}{\eta_{b,out}}, & n > \frac{P_{storage}}{P_{g1}} \end{cases} \quad (3-14)$$
Assume a value N, when n = N, F(n) is the maximum value of the function. The maximum overall efficiency of the storage system is:
$$\eta_t = \frac{F(N)}{P_{storage}} \times 100\% \quad (3-15)$$
When the overall system is in the charging state, the system can be divided into four typical modes for steady-state analysis.
State 1: Gravity storage system and battery are both charging (Pg > 0, Pb > 0). The efficiency formula is:
$$F_1(n) = \frac{n \times P_{g1} \times \eta_{g,in} + P_{storage} \times \eta_{b,in}}{P_{storage}} \quad (3-16)$$
State 2: Only gravity storage system is charging, battery is inactive (Pg > 0, Pb = 0). The efficiency formula is:
$$F_2(n) = \frac{n \times P_{g1} \times \eta_{g,in}}{P_{storage}} \quad (3-17)$$
State 3: Gravity storage system is inactive, battery is charging (Pg = 0, Pb > 0). The efficiency formula is:
$$F_3(n) = \eta_{b,in} \times P_{storage} \quad (3-18)$$
State 4: Gravity storage system is charging, battery is discharging (Pg > 0, Pb < 0). The efficiency formula is:
$$F_4(n) = \frac{n \times P_{g1} \times \eta_{g,in} + P_{storage} \times \eta_{b,out}}{P_{storage}} \quad (3-19)$$
For the discharge state, when the storage system is discharging:
$$P_{storage} < 0 \quad (3-20)$$
$$P_{storage} = P_g + P_b \quad (3-21)$$
The discharge efficiency can be expressed as:
$$\eta = \frac{P_{storage}}{P_{g,out} + P_{b,out} + P_{b,in}} \quad (3-22)$$
Since a smaller denominator means higher efficiency when Pstorage is externally determined, the function relating to n and Pb is:
$$T(n, P_b) = \begin{cases} n \times P_{g1} + P_b \times \eta_{b,out}, & P_b > 0 \\ n \times P_{g1} + \frac{P_b}{\eta_{b,in}}, & P_b < 0 \end{cases} \quad (3-23)$$
Similarly, substituting Pb = Pstorage – n × Pg1 yields the simplified function T(n):
$$T(n) = \begin{cases} n \times P_{g1} \times \eta_{g,out} + (P_{storage} – n \times P_{g1}) \times \eta_{b,out}, & n < \frac{P_{storage}}{P_{g1}} \\ \frac{n \times P_{g1}}{\eta_{g,in}} + \frac{P_{storage} – n \times P_{g1}}{\eta_{b,in}}, & n > \frac{P_{storage}}{P_{g1}} \end{cases} \quad (3-24)$$
Likewise, the discharge state can be divided into four states:
State 1: Both gravity storage system and battery are discharging (Pg < 0, Pb < 0).
State 2: Gravity storage system is discharging, battery is inactive (Pg < 0, Pb = 0).
State 3: Gravity storage system is inactive, battery is discharging (Pg = 0, Pb < 0). The efficiency equals ηb,out.
State 4: Gravity storage system is discharging, battery is charging (Pg < 0, Pb > 0).
3.2 Weight Allocation Based on Analytic Hierarchy Process
Set the weights of maximum efficiency, battery remaining capacity, load demand condition, and new energy generation condition as A, B, C, and D respectively. The judgment matrix is shown in the table below.
| Maximum Efficiency | Battery Remaining Capacity | Load Demand | New Energy Generation | |
|---|---|---|---|---|
| Maximum Efficiency | 1 | A/B | A/C | A/D |
| Battery Remaining Capacity | B/A | 1 | B/C | B/D |
| Load Demand | C/A | C/B | 1 | C/D |
| New Energy Generation | D/A | D/B | D/C | 1 |
From the table, the eigenvectors of maximum efficiency, battery remaining capacity, load demand condition, and new energy generation condition can be obtained as:
$$a_{max} = \sqrt[4]{1 \times \frac{A}{B} \times \frac{A}{C} \times \frac{A}{D}} \quad (3-25)$$
$$a_{battery} = \sqrt[4]{\frac{B}{A} \times 1 \times \frac{B}{C} \times \frac{B}{D}} \quad (3-26)$$
$$a_{load} = \sqrt[4]{\frac{C}{A} \times \frac{C}{B} \times 1 \times \frac{C}{D}} \quad (3-27)$$
$$a_{new} = \sqrt[4]{\frac{D}{A} \times \frac{D}{B} \times \frac{D}{C} \times 1} \quad (3-28)$$
After normalization:
$$\omega_{max} = \frac{a_{max}}{a_{max} + a_{battery} + a_{load} + a_{new}} \quad (3-29)$$
$$\omega_{battery} = \frac{a_{battery}}{a_{max} + a_{battery} + a_{load} + a_{new}} \quad (3-30)$$
$$\omega_{load} = \frac{a_{load}}{a_{max} + a_{battery} + a_{load} + a_{new}} \quad (3-31)$$
$$\omega_{new} = \frac{a_{new}}{a_{max} + a_{battery} + a_{load} + a_{new}} \quad (3-32)$$
Using the square root method, solving for each criterion under the four states yields the normalized eigenvectors. The combined weight of each state is then calculated as:
$$A_{weight} = \omega_{max} \times A_{state1} + \omega_{battery} \times B_{state1} + \omega_{load} \times C_{state1} + \omega_{new} \times D_{state1} \quad (3-33)$$
$$B_{weight} = \omega_{max} \times A_{state2} + \omega_{battery} \times B_{state2} + \omega_{load} \times C_{state2} + \omega_{new} \times D_{state2} \quad (3-34)$$
$$C_{weight} = \omega_{max} \times A_{state3} + \omega_{battery} \times B_{state3} + \omega_{load} \times C_{state3} + \omega_{new} \times D_{state3} \quad (3-35)$$
$$D_{weight} = \omega_{max} \times A_{state4} + \omega_{battery} \times B_{state4} + \omega_{load} \times C_{state4} + \omega_{new} \times D_{state4} \quad (3-36)$$
Based on the weights of each state, the current optimal operating state can be determined. In the charging and discharging efficiency optimization model of the hybrid storage system, the weight distribution follows these principles: system efficiency always occupies the highest priority (0.8 in this paper). Battery remaining capacity, as a dynamic variable, directly affects charging and discharging efficiency, so its weight coefficient needs to be adaptively adjusted in real time with the SOC state.
For the load demand parameter, under charging conditions, although the weight needs to be dynamically corrected with load fluctuations, its influence coefficient should be controlled within a low threshold range. For the new energy generation power parameter, its weight configuration needs to consider both the current operating state and its correlation with subsequent discharge strategies. Its weight value not only needs to be higher than the load demand parameter but should also establish a nonlinear mapping relationship based on real-time input power fluctuations.
4. Comparative Study of Gravity-Battery Hybrid Storage Optimization
This chapter mainly simulates the comprehensive efficiency of the hybrid storage system under different operating conditions and the respective handling of gravity storage and battery in the hybrid storage system, based on the temperature changes of a typical day. Combined with the optimization strategy proposed in Chapter 3, the feasibility of the optimization strategy is verified through simulation.
4.1 Temperature Impact Analysis
Temperature and battery remaining capacity changes have a significant impact on the charging and discharging efficiency of the battery-gravity hybrid system. The mechanisms can be analyzed from multiple dimensions.
(1) Temperature sensitivity of the battery system. When the battery operates in a low-temperature environment below 10°C, the battery faces capacity attenuation problems. The low temperature increases electrolyte viscosity, hindering lithium-ion diffusion and reducing the available capacity of the battery. The charging and discharging efficiency also decreases. The cycle life of the battery is also affected, with charging in low-temperature environments easily triggering lithium plating or plate crystallization.
When the battery is in a high-temperature environment above 35°C, the battery faces problems of capacity false increase and life attenuation. High temperatures accelerate chemical reactions, temporarily increasing capacity, but long-term high-temperature operation causes electrolyte decomposition and SEI film thickening. High-temperature environments also pose thermal runaway risks, and self-discharge effects intensify.
(2) Temperature adaptability advantage of gravity storage. The gravity energy storage system has mechanical stability and is less affected by external environments. Since the charging and discharging of gravity storage depends on potential energy conversion of heavy blocks, the mechanical transmission efficiency is little affected by temperature, fluctuating little in the range of -30°C to 50°C. The low thermal expansion coefficient of heavy blocks makes the impact of deformation on track friction loss negligible under all but extreme temperatures.
(3) Temperature synergistic optimization mechanism of the hybrid system. During high-temperature periods, gravity storage undertakes base load regulation (4-12 hours of continuous output), reducing battery high-load operation time and avoiding temperature accumulation. During low-temperature periods, the battery operates in shallow charge/discharge mode to handle second-level fluctuations, while gravity storage provides more than 80% of the energy support, reducing the battery’s low-temperature attenuation risk.
The efficiency of lithium batteries is affected by temperature and can be modeled based on the Arrhenius equation:
$$\eta(T) = \eta_{ref} \exp\left(\frac{E_a}{R}\left(\frac{1}{T_{ref}} – \frac{1}{T}\right)\right) \quad (4-1)$$
where Ea is the reaction activation energy (J/mol); R is the ideal gas constant (8.314 J/(mol·K)); T is the battery temperature (K); Tref is the reference temperature (298 K, i.e., 25°C); and ηref is the nominal efficiency at the reference temperature (95%).
In practical engineering, a piecewise linear approximation is often used (applicable to 20°C-60°C):
$$\eta(T) = \begin{cases} \eta_{ref} [1 – k_1 (T_{ref} – T)], & T < T_{ref} \\ \eta_{ref} [1 – k_2 (T – T_{ref})], & T > T_{ref} \end{cases} \quad (4-2)$$
where k1 is the low-temperature efficiency decay coefficient and k2 is the high-temperature efficiency decay coefficient.
Based on the above formulas, the battery temperature-efficiency relationship is shown in the table below.
| Temperature Range | Efficiency Correction Factor | Typical Efficiency Range (0.5C) |
|---|---|---|
| -20°C | η reduced | 50%-60%, significant capacity decay |
| 0°C-10°C | η reduced | 85%-90%, preheating needed |
| 25°C | 95%-98% | Optimal operating range |
| 40°C-60°C | η reduced | 90%-93%, side reactions accelerate |
4.2 Charging and Discharging Efficiency Comparison in Different Scenarios
This paper adopts wind power and photovoltaic power generation data from a certain new energy generation company in Guangdong Province in January, combined with the day’s temperature for simulation verification. The simulation uses one day of data from a certain region in Guangdong Province as the load demand. According to the simulation results of the battery efficiency in a typical day in January, the battery operating efficiency presents a fluctuation range of 82.72%-91.74%.
Based on the load demand of a certain area in Guangdong Province and the power input conditions of different new energy power plants, the system simulation is conducted based on the energy allocation strategy proposed in Chapter 3 (without considering battery remaining capacity weight). The simulation results show that the hybrid energy storage system is superior to a single gravity storage or battery storage system in charging and discharging efficiency. In the charging state, the comprehensive efficiency of the hybrid system improves by 3.74% compared to the single battery system, and by 3.89% compared to the gravity storage system. In the discharging state, the hybrid system efficiency improves by 4.21% compared to the single battery system, and by 4.25% compared to the gravity storage system.
After adopting the structural optimization method proposed in Chapter 2 to improve the fitting degree of the gravity storage output power to the load demand by 37%, simulation is conducted by combining the hybrid storage optimization method proposed in Chapter 3. The simulation results show that when the gravity storage fitting degree increases by 37%, the comprehensive efficiency of the hybrid energy storage system improves by 2.25%, 3.21%, and 3.54% respectively under the three new energy input scenarios of photovoltaic power station, wind farm No. 1, and wind farm No. 2. In this state, the equivalent gravity storage efficiency improvement is about 4%.
| Scenario | Efficiency Improvement After 37% Fit Increase | Equivalent Gravity Storage Efficiency Gain |
|---|---|---|
| Photovoltaic power station | 2.25% | ~4% |
| Wind farm No. 1 | 3.21% | ~4% |
| Wind farm No. 2 | 3.54% | ~4% |
In practical engineering, compared to the ultra-large capacity (GWh-level) of gravity storage, the capacity of batteries (kWh-MWh level) is significantly smaller. Therefore, when calculating the efficiency of the hybrid storage system, the effects of battery SOC and charge/discharge rate C need to be considered:
$$\eta(T, SOC, C) = \eta(T) [1 – \alpha(1 – SOC)] \exp(-\beta \cdot C) \quad (4-3)$$
where α is the SOC correction coefficient; β is the rate correction coefficient; and C is the charge/discharge rate.
After considering the battery remaining capacity, the impact of SOC on battery charge/discharge efficiency and the impact of SOC on weights in Chapter 3 are calculated. Simulation verification is conducted with a battery capacity of 1 MWh and an initial SOC of 40%. The simulation results show that the battery output curve exhibits significant dual-timescale characteristics. When the SOC is low and the system is in the charging period, the battery prioritizes charging to increase its remaining capacity. When the remaining capacity reaches a certain level, the battery stops charging due to reduced efficiency, and the gravity storage system stores all the remaining electricity. In the discharge phase, when the remaining battery capacity is high and its efficiency is higher than that of gravity storage, the system allows the battery to discharge alone or jointly with gravity storage to improve efficiency. As the remaining battery capacity decreases, the proportion of gravity storage discharge gradually increases.
After calculating SOC, the battery efficiency decreases by about 3.24% compared to the case without SOC calculation, with a daily comprehensive efficiency of about 84.04%. The charging and discharging efficiency of the hybrid storage system on that day is obtained by combining the optimization method of Chapter 3, achieving a comprehensive efficiency of 87.13%, which is 3.39% higher than that of the single battery storage system and 2.43% higher than that of the independent gravity storage system (baseline efficiency 85.00%).
| System Configuration | Comprehensive Efficiency | Improvement Over Battery-Only | Improvement Over Gravity-Only |
|---|---|---|---|
| Battery-only (with SOC effect) | 84.04% | — | — |
| Gravity-only | 85.00% | — | — |
| Hybrid system | 87.13% | +3.39% | +2.43% |
From the efficiency distribution of the hybrid storage system at typical operating points, a bimodal characteristic is observed. In the charging phase’s SOC rising region, the peak system efficiency reaches 88.92%, mainly benefiting from the high-efficiency characteristics of the battery during constant current charging. In the discharging phase’s SOC falling region, the efficiency curve exhibits a wide plateau, mainly due to the gravity storage system smoothing the gradient change of battery efficiency as SOC decreases.
The simulation results validate the effectiveness of the hybrid storage optimization method proposed in Chapter 3. The advantages of the hybrid storage system are demonstrated in two aspects: on the one hand, the battery can undertake transient power fluctuation regulation tasks in the high-efficiency region of SOC, fully utilizing the fast response advantage of the battery. On the other hand, the gravity storage system can provide basic power support through its stable efficiency in the boundary region of battery SOC.
In the simulation where the capacity ratio of gravity storage to battery is configured, the gravity storage system in the hybrid storage system is responsible for long-period base load regulation tasks, effectively reducing the battery cycle depth. The battery is only responsible for peak power regulation within a short period, greatly extending the battery’s service life. This capacity allocation method takes full advantage of the gravity storage system’s large capacity and long lifespan characteristics, while making up for the shortcomings of the battery’s limited cycle life. According to estimates, the annual average cycle count of lithium batteries can be reduced by more than 60% after adopting this hybrid configuration.
The entire simulation verification process is based on the MATLAB/Simulink platform. By building the gravity-battery hybrid storage charging and discharging model and the hybrid storage energy allocation optimization model, simulations under different new energy generation conditions are conducted. The results show that the proposed hybrid storage strategy improves the overall efficiency of the hybrid energy storage system under different operating scenarios.
5. Conclusion and Outlook
This paper focuses on the characteristics of the gravity energy storage system itself and the collaborative operation mechanism of hybrid storage. Through multi-objective collaborative optimization methods, the deep coupling of the long-period peak shaving of gravity storage and the fast power support of batteries is achieved, providing a storage solution with both economy and reliability for the construction of the new power system. The main research conclusions are as follows:
(1) Based on the physical characteristics of gravity storage itself, a dynamic energy state model considering mechanical transmission efficiency and potential energy conversion loss was constructed. An innovative “auxiliary storage yard – segmented slope – intelligent reversing” three-dimensional structural optimization framework was proposed. Through establishing modular storage yard layouts, multi-gradient potential energy conversion channels, and adaptive track control systems, the system response time was greatly reduced while construction costs were lowered. The fitting degree between gravity storage output power and load demand was improved, thereby enhancing the comprehensive efficiency of the hybrid storage system and significantly strengthening the dynamic regulation capability of the gravity storage system in response to power fluctuations.
(2) To address the difficulty of collaborative operation of heterogeneous storage units, a composite weight evaluation system including full lifecycle cost, dynamic efficiency, new energy generation conditions, and load demand conditions was constructed. The weight allocation strategy was established based on the analytic hierarchy process. By switching the operating states within the hybrid storage system, the comprehensive efficiency of the hybrid storage system was improved.
(3) Based on the MATLAB/Simulink platform, the charging and discharging model of the gravity-battery hybrid storage system and the hybrid storage energy allocation optimization model were built. Simulation verification was conducted for different new energy generation scenarios. The simulation results show that the proposed hybrid storage strategy improves the overall efficiency of the hybrid storage system under different operating scenarios.
Future research will focus on breaking through the geographical adaptability assessment model of gravity storage and exploring the multi-energy flow coordinated control mechanism with compressed air storage, hydrogen storage, and other multi-type storage technologies. Specifically, the dynamics optimization of the gravity storage system can be further improved by studying the collaborative optimization of mechanical transmission systems such as dynamic slope angle adjustment and composite track friction coefficient optimization. The research in this paper is based on empirical analysis of typical winter conditions in Guangdong Province; due to the stability of the winter temperature range in Guangdong, the thermodynamic characteristics of the storage system were not fully excited, resulting in little change in temperature-sensitive efficiency indicators. Future research should expand to typical extreme temperature gradient scenarios such as northwest high-altitude or continental climate regions, focusing on constructing a temperature-efficiency coupling model to quantify the synergistic impact mechanism of gravity storage and battery characteristics on hybrid system performance under extreme temperature differences.